Upload
trinhnhu
View
221
Download
0
Embed Size (px)
Citation preview
XML-IS
Our reference: NLM 1790 P-authorquery-vx
AUTHOR QUERY FORM
Journal: NLM
Please e-mail or fax your responses and any corrections to:
Article Number: 1790
E-mail: [email protected]
Fax: +44 1392 285878
Dear Author,
Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screenannotation in the PDF file) or compile them in a separate list.
For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.
Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in the proof.
Click on the Q link to go to the location in the proof.
Location inarticle
Query / Remark: click on the Q link to go
Please insert your reply or correction at the corresponding line in the proof
Q1 Please check whether the designated corresponding author is correct, and amend if necessary.
Q2 The spelling of the author name(s) in the text has been changed to ‘‘Grandhi and Wolons [10]’’ to match thereference list. Please check the spelling, and correct if necessary.
Thank you for your assistance.
Q1
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
63
65
International Journal of Non-Linear Mechanics ] (]]]]) ]]]–]]]
Contents lists available at ScienceDirect
International Journal of Non-Linear Mechanics
0020-74
doi:10.1
� Corr
E-m
Pleasstay
journal homepage: www.elsevier.com/locate/nlm
Performance evaluation of shape-memory-alloy superelastic behavior tocontrol a stay cable in cable-stayed bridges
O. Ben Mekki a,�, F. Auricchio b
a University of Tunis EL Manar, ENIT, Laboratory of Civil Engineering (LGC), Tunisiab Department of Structural Mechanics, University of Pavia, Pavia, Italy
a r t i c l e i n f o
Article history:
Received 8 July 2009
Received in revised form
23 November 2010
Accepted 1 December 2010
Keywords:
Shape-memory alloys
Stay cable
Superelastic
Dampers
62/$ - see front matter & 2010 Elsevier Ltd. A
016/j.ijnonlinmec.2010.12.002
esponding author.
ail address: [email protected] (O. Ben Mekki)
e cite this article as: O. Ben Mekki, F.cable in cable-stayed bridges, Int. J.
a b s t r a c t
This paper focuses on introducing and investigating the performance of a new passive control device for
stay cable in cable-stayed bridges made with shape-memory alloys (SMA). The superelasticity and
damping capability of SMA is sought in this study to develop a supplementary energy dissipation device
for stay cable. A linear model of a sag cable and a one-dimensional constitutive model for the SMA are
used. The problem of the optimal design of the device is studied. In the optimization problem, an energy
criterion associated with the concept of optimal performance of the hysteretic connection is used. The
maximum dissipation energy depends on the cross-sectional area, the length, and the location of the SMA
on the cable. The effectiveness of the SMA damper in controlling the cable displacement is assessed.
Furthermore, a study is conducted to determine the sensitivity of the cable response to the properties of
the SMA device. The comparison between the SMA damper and a more classical passive control energy
dissipation device, i.e., the tuned mass damper (TMD), is carried out. The numerical results show the
effectiveness of the SMA damper to damp the high free vibration and the harmonic vibration better than
an optimal TMD.
& 2010 Elsevier Ltd. All rights reserved.
67
69
71
73
75
77
79
81
83
85
87
89
1. Introduction
Over the last few decades, cable-stayed bridges have attractedgreat interest because of their aesthetics, structural efficiency, andeconomy. This type of construction has become popular worldwidein recent years, largely due to the rapid progress in designmethodology and construction technologies [26]. However, staycables are critical structural components in these bridges. Owing totheir large flexibility, relatively small mass and extremely lowdamping, stay cables have frequently exhibited large-amplitudevibrations under wind, wind–rain and support motion. Aerody-namic instability of stay cables with extremely large oscillationamplitude under specific rain and wind conditions has beenobserved in a number of cable-stayed bridges worldwide, and itis a conundrum to civil engineers [29,39]. Therefore, the mitigationof dynamic response quantities induced by environmental loads isof vital importance in terms of safety and serviceability [2,25].
In the past decade, cable vibration control techniques by meansof passive countermeasures, including aerodynamic, mecha-nical and structural means, have been broadly investigated andsuccessfully implemented [17]. In the meantime, researchers havealso studied the active vibration control of cables by applying
91
93
ll rights reserved.
.
Auricchio, Performance evaNon-Linear Mech. (2010), d
transverse force control and axial stiffness or tension (supportmotion) control [30,35].
A lot of researches have been conducted to investigate possibledamping systems and to determine the optimal size of viscousdampers attached to cables for vibration control. Kovacs [15] wasamong the first to investigate the maximum attainable damping ratiofor a taut cable with a viscous damper. Pachero et al. [3] proposed a‘‘universal estimation curve’’ of normalized modal damping ratiosversus normalized damper coefficient for a horizontal taut stringmodel. This ‘‘universal estimation curve’’ is generalized by Cremona[5] for inclined cables by taking account of the sag-extensibilityparameter. A transfer matrix formulation is developed by Xu et al.[40] to estimate the modal damping ratio of inclined cables attachedwith oil dampers. Main et al. [18] proposed an analytical formulationof a taut cable with an attached damper. Theoretical studies were alsocarried out to evaluate the increased damping level of a stay cableafter installing passive viscous dampers [2]. It was found that thereexists an optimum viscous coefficient of the damper by which themodal damping ratio of a stay cable can reach its maximum for agiven mode of vibration. However, this passive device suffers fromseveral drawbacks such as the modal damping ratio of the stay cabledecreases rapidly when the viscous coefficient deviates from itsoptimal value. The use of a variable-orifice viscous damper andelectrorheological or magnetorheological (ER/MR) fluid damper withsemi-active control may be an alternative [38]. However, the semi-active control device is more complicated to implement. Another
95
luation of shape-memory-alloy superelastic behavior to control aoi:10.1016/j.ijnonlinmec.2010.12.002
Q2
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
95
97
99
101
103
105
107
109
111
113
115
119
121
123
125
127
129
131
133Fig. 1. Schematic diagram of an inclined stay cable.
O. Ben Mekki, F. Auricchio / International Journal of Non-Linear Mechanics ] (]]]]) ]]]–]]]2
candidate that has a great potential for vibration control of stay cablesubjected to wind and wind/rain loads is a superelastic shape-memory alloy (SMA) damper with the advantages of large dampingcapacity, self-centering ability, high fatigue-resistant performanceand good corrosion resistance [21].
As a natural consequence of the microscopic behavior, at themacroscopic level shape-memory solids present the superelasticeffect (the recovery of large deformations in loading–unloadingcycles, occurring at sufficiently high temperatures) and the shape-memory effect (the recovery of large deformations by a combina-tion of mechanical and thermal processes).
These unique properties enable SMA to be used as actuators,passive energy dissipators and dampers for civil structural control[14,1,11,32,31]. When integrated with civil structures, SMAs can bepassive, semi-active, or active components to reduce damagecaused by environmental impacts.
Using SMAs for passive structure control relies on the SMA’sdamping capacity, which represents its ability to dissipate vibra-tion energy of structures subject to dynamic loading.
Several authors investigated the energy dissipation of widelyused Nitinol superelastic SMA wires. Dolce and Cardone [23]studied superelastic Nitinol wires subjected to tension loading.They observed the dependence of the damping capacity ontemperature, loading frequency and the number of loading cycles.Grandhi and Wolons [10] proposed using a complex modulusapproach to characterize the damping capacity of superelastic SMAwires for convenient integration with structure dynamics. A super-elastic SMA wire demonstrates the damping capacity not onlyunder tension loading, but also under cyclic bending [20]. Thenumerical results showed that the energy dissipated by thesuperelastic SMA wire is highly sensitive to its diameter, i.e., thethicker the SMA wire, more energy is dissipated.
SMA energy dissipation devices have been seen in the forms ofbraces for framed structures [37], connection elements for columns[16], retrofitting devices for historic building [24] and dampers forsimply supported bridges [28].
Recently, as large cross-sectional area SMA elements are becom-ing available, and studies on the properties of SMA bars haveattracted more attentions [23,36]. As indicated in Ref. [36], thedamping capacity of a martensite Nitinol bar under tension–compression cycles increases with increasing strain amplitude,but decreases with loading cycles and then reaches a stableminimum value. The optimization of the cross-sectional area andthe length of the SMA device is presented in [22]. The dynamicperformance of the device is evaluated by the steady-state responseat the resonance point in order to focus on the damping effect.Analytical formulation utilizing the equivalent linearizationapproach successfully leads to the basic correlation between thehysteresis shape and the damping effect.
To explore the potentials of SMA based energy dissipation inpassive structure control, this paper presents an approach to studythe damping vibration of stay cables in a cable-stayed bridge byusing a SMA energy dissipation devices with superelastic hyster-esis. The first part of the paper presents the general three-dimensional equations of a stay cable subjected to externaldynamic loading and controlled by a distribution of dampers inthe transverse direction, detailing the hypothesis of the problemlinearization. The second part of the paper focuses on three aspects:formulating a mathematical model of the cable with one SMAdamper using a Galerkin approximation, verifying the feasibility ofthe SMA to control the stay cable, and optimizing the SMA deviceusing numerical method. For this simulation a one-dimensionalmodel for superelastic SMA [9] is considered. Finally, the third partof the paper, focuses on the comparison between the TMD and theSMA energy dissipation device to control the stay cable free andharmonic transverse vibration.
Please cite this article as: O. Ben Mekki, F. Auricchio, Performance evastay cable in cable-stayed bridges, Int. J. Non-Linear Mech. (2010), d
2. Dynamic equations formulation of a sag stay cable
A cable is a spatially distributed system, whose transversaldimensions are significantly smaller than longitudinal one. Staycables have very low levels of inherent mechanical damping andthe mechanisms associated with the observed large-amplitudevibrations are still not fully understood. This section has two goals;the first goal is to present the general three-dimensional equationsoff a stay cable subjected to external dynamic loading andcontrolled by a distribution of dampers in the transverse direction.The second goal is to introduce the hypothesis leading to theproblem linearization; i.e., the assumption of a small sag and theuncoupling of the in-plane and out-of-plane behaviors.
We start considering a cable connecting two points denoted as A
and B and placed to a distance L. The segment connecting A and B
defines an angle y versus horizontal axis. For the case of only bodyload, the cable configuration is planar and we indicate with Axy anorthogonal reference system defined within such a plane. Theplanar oscillations can occur in the transverse direction (y-axis) asshown in Fig. 1 and non-planar oscillations can occur in directionAz, so that Axyz forms a direct orthogonal frame.
The equations governing the static equilibrium of an inclinedcable element subjected only to dead load (gravity) are
@
@sT
dx
ds
� �¼mgsiny
@
@sT
dy
ds
� �¼�mgcosy
8>>><>>>:
ð1Þ
where s is the Lagrangian co-ordinate, T is the static cable tension,m is the mass of the cable per unit length, and g is the accelerationdue to gravity.
A non-linear dynamic model of an inclined cable is built in thecoordinate system (Axyz) by three coupled partial differentialequations [13]:
@
@sðTþtÞ dx
dsþ@u
@s
� �� �þFxðx,tÞ ¼m
@2u
@t2þmgsiny
@
@sðTþtÞ dy
dsþ@v
@s
� �� �þFyðx,tÞ�
XMi ¼ 1
fc,iðtÞdðs�sc,iÞ ¼m@2v
@t2�mgcosy
@
@sðTþtÞ@w
@s
� �þFzðx,tÞ ¼m
@2w
@t2
8>>>>>>>>><>>>>>>>>>:
ð2Þ
where t is the dynamic cable tension; u, v and w are the cabledynamic displacement components in the x-, y- and z-directions,respectively, measured from the position of the static equilibrium
luation of shape-memory-alloy superelastic behavior to control aoi:10.1016/j.ijnonlinmec.2010.12.002
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
95
97
99
101
103
105
107
109
111
113
115
119
121
123
125
127
129
131
133Fig. 2. In-plane stay cable with sag attached with one transverse SMA.
O. Ben Mekki, F. Auricchio / International Journal of Non-Linear Mechanics ] (]]]]) ]]]–]]] 3
of the cable; Fx, Fy and Fz are distributed external dynamic loadingper unit length in the x-, y- and z-directions, respectively; fc, i areconcentrated external force applied by the ith external damper onthe cable at the location sc, i in the transverse direction; sc, i is theLagrangian co-ordinate of the ith dampers measured from the leftsupport of the cable; t is the time; M is the total number of thedampers; dð�Þ is the dirac’s delta function.
We consider the following boundary conditions for the cable:
uð0,tÞ ¼ vð0,tÞ ¼wð0,tÞ ¼ 0
uðL,tÞ ¼ vðL,tÞ ¼wðL,tÞ ¼ 0
(ð3Þ
2.1. Equation of motion for a small sag cable
In a cable-stayed bridge, the stay cables are generally stronglypre-constrained. Thus, the sag is assumed to be small, ds� dx, andthe static tension in the cable is assumed to be constant along thespan. Due to these two hypotheses, the static shape of an ‘‘inclinedcable’’ is then close to the chord. Accordingly, it is reasonable toapproximate this deformation by a parabola whose equation isdefined in the local coordinate system (Axyz) as follows [4]:
yðxÞ ¼ 4dx
L�
x
L
� �2� �
ð4Þ
where d¼mgL2cosy=8T is the sag at mid-span.Using Eqs. (1) governing the static equilibrium and introducing
the hypothesis of a small sag cable, system (2) is simplified:
@
@xðTþtÞ@u
@xþt
� �þFxðx,tÞ ¼m
@2u
@t2
@
@xðTþtÞ@v
@xþtdy
dx
� �þFyðx,tÞ�
XMi ¼ 1
fc,iðtÞdðx�xc,iÞ ¼m@2v
@t2
@
@xðTþtÞ@w
@x
� �þFzðx,tÞ ¼m
@2w
@t2
8>>>>>>>>><>>>>>>>>>:
ð5Þ
Finally, it is assumed that the dynamic tension of the cable is relatedto the non-linear dynamic strain by the following relation [4]:
tðx,tÞ ¼ EAeðx,tÞ ð6Þ
where E is Young’s modulus, A is the constant cross-sectional area ofthe cable, and eðx,tÞ is the non-linear axial strain in the cable definedas follows [6]:
eðx,tÞ � eðtÞ ¼ @u
@xþ
dy
dx
@v
@xþ
1
2
@u
@x
� �2
þ@v
@x
� �2
þ@w
@x
� �2" #
ð7Þ
2.2. Non-coupled model of small sag stay cable
For a small sag, the cable behaves as a chord; its transversalfrequency is smaller than its longitudinal frequency. Therefore, thelongitudinal inertia force m@2u=@t2 is assumed to be negligible, andthe longitudinal deformation to be small [6]. When considering thelongitudinal external forces to be zero, the first equation of system(5) shows that the dynamic tension in the cable can be assumedconstant along the cable span. Thus, introducing this result in thetwo other Eqs. (5), we get
@tðx,tÞ
@x¼ 0
ðTþtÞ@2v
@x2þtd2y
dx2þFyðx,tÞ�
XMi ¼ 1
fc,iðtÞdðx�xc,iÞ ¼m@2v
@t2
ðTþtÞ@2w
@x2þFz ¼m
@2w
@t2
8>>>>>>>><>>>>>>>>:
ð8Þ
In order to linearize the previous equations of the small sag cable,the following hypothesis are assumed: small deformations which
Please cite this article as: O. Ben Mekki, F. Auricchio, Performance evastay cable in cable-stayed bridges, Int. J. Non-Linear Mech. (2010), d
implies that second order terms can be neglected, and the dynamictension is negligible compared to be the static tension. Therefore,the dynamic cable tension can be defined by
tðtÞ ¼ EA@u
@xþ
dy
dx
@v
@x
� �, ð9Þ
and the last two equations of system (8) are written as
T@2v
@x2þtd2y
dx2þFyðx,tÞ�
XMi ¼ 1
fc,iðtÞdðx�xc,iÞ ¼m@2v
@t2
T@2w
@x2¼m
@2w
@t2
8>>>><>>>>:
ð10Þ
The linear theory applied to a cable with a small sag shows thatin-plane and out-of-plane behaviors are essentially uncoupledbecause this motion involves no additional cable tension (to thefirst order); the out-of-plane modes and the in-plane antisym-metric modes are the same as those of a taut string, while thein-plane symmetric modes are controlled by the Irvine coefficient.
Using Eq. (9), the longitudinal displacement is deducted fromthe first equation of system (8), i.e.:
@2u
@x2¼�
@
@x
dy
dx
@v
@x
� �ð11Þ
By double-integrating equation (11) and introducing the boundaryconditions as indicated by system (3), the dynamic tension is given by
tðtÞ ¼� EA
L
d2y
dx2
Z L
0vðx,tÞ dx ð12Þ
By replacing Eq. (12) in the first equation of system (10), thetransversal displacement of the stay cable verifies this linear equa-tion:
m@2v
@t2�T
@2v
@x2þ
EA
L
d2y
dx2
� �2 Z L
0vðx,tÞ dx¼ Fyðx,tÞ�
XMi ¼ 1
fc,iðtÞdðx�xc,iÞ
ð13Þ
where d2y=dx2 and dy=dx can be deducted from Eq. (4).This linear model is obtained according to two main hypoth-
eses; the sag cable is small and the in-plane and out-of-planebehaviors are essentially uncoupled. This last model is tackled tostudy the control of the stay cable considering one SMA damperattached near the cable’s lower end.
3. Control of stay cable transverse vibration with one SMAdamper
In this section, we consider a stay cable suspended between twosupports and equipped by a SMA damper installed at a distance xc
from support A as shown in Fig. 2. Generally, the damper locationon the stay cable has to be near a support because it is the mostpractical position. We assume that the SMA damper produces aforce fc in the y-direction. Considering the end-supports of the cable
luation of shape-memory-alloy superelastic behavior to control aoi:10.1016/j.ijnonlinmec.2010.12.002
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
95
97
99
101
103
105
107
109
111
113
115
119
121
123
125
127
129
131
133
O. Ben Mekki, F. Auricchio / International Journal of Non-Linear Mechanics ] (]]]]) ]]]–]]]4
to be fixed, the combined stay cable/SMA system has the followingpartial differential equation of motion derived from Eq. (13):
m €v�Tv00 þEA
Lðy00Þ2
Z L
0v dx¼ Fy�fcðtÞdðx�xcÞ ð14Þ
where we now use notation, ðuÞ and ð_Þ to denote partial derivativeswith respect to x and t, respectively, and where we recall that
v¼ vðx,tÞ, Fy ¼ Fyðx,tÞ, and y¼ yðxÞ ð15Þ
3.1. Approximate series solutions to equations of motion
The transverse deflexion can be approximated using a finite series
v¼XN
i ¼ 1
aiðtÞjiðxÞ ð16Þ
whereaiðtÞ are non-dimensional modal participation factors andjiðxÞ
are a set of mode shape functions, assumed to be continuous and tosatisfy the geometric boundary conditions, i.e., jið0Þ ¼ jiðLÞ ¼ 0.
To compute the damping and responses of the cable with theSMA damper, we assume sinusoidal shape functions, jiðxÞ ¼
sinðipx=LÞ, identical to the mode shapes of the cable without thedamper. Adopting Galerkin approximation based on the free modesof the stay cable, the factors aiðtÞ satisfy
mii €ai ðtÞþðkiiþl2miiÞaiðtÞ ¼ Fyi�fcðtÞjiðxcÞ for i¼ 1 . . .N ð17Þ
where
l2¼ EA
Lmgcosy
T
� �2
mii ¼mR L
0 jiðxÞjiðxÞ dx¼mL2
kii ¼ TR L
0 juiðxÞjuiðxÞ dx¼ Tp2i2
2L
Fyi ¼R L
0 Fyðx,tÞjiðxÞ dx
jiðxcÞ ¼ sin ipxc
L
� �
8>>>>>>>>>><>>>>>>>>>>:
ð18Þ
We observe that the system of dynamical equations (17) are decoupledin all terms except than the term fc(t) induced by the presence of theSMA damper. So the SMA damper introduces non-linearity intocombined linear cable-damper system and Newmark numericalmethod is used to compute the dynamic response of the cable.
In the following we detail the expression of the damping forcefc(t), which will have a form clearly depending on the SMAconstitutive behavior.
3.2. Constitutive model for the SMA system
To reproduce the superelastic behavior of the SMA elementintroduced in the previous section, we consider the one-dimen-sional model proposed in Ref. [9]. This model considers one scalarinternal variable, xs, representing the martensite fraction, and twoprocesses which may produce martensite fraction variations: theconversion of austenite into martensite (A�!S), and the conversionof martensite into austenite (S�!A).
Following experimental evidence, both the processes are gov-erned by linear kinetic rules in terms of the uniaxial stress s. Inparticular, the activation conditions for the conversion of austeniteinto martensite are
sASs o jsjosAS
f and _jsj40 ð19Þ
wheresASs andsAS
f are material parameters, j � j is the absolute valueand a superpose dot indicates a time-derivative. The correspondingevolution equation is set equal to
_xs ¼�ð1�xsÞ_jsj
jsj�sASf
ð20Þ
Please cite this article as: O. Ben Mekki, F. Auricchio, Performance evastay cable in cable-stayed bridges, Int. J. Non-Linear Mech. (2010), d
On the other hand, the activation conditions for the conversion ofmartensite into austenite are
sSAf o jsjosSA
s and _jsjo0 ð21Þ
where sSAf and sSA
s are material parameters. The correspondingevolution equation is set equal to
_xs ¼ xs
_jsjjsj�sSA
f
ð22Þ
Limiting the discussion to a small deformation regime, the modelproposed in Ref. [9] assumes an additive decomposition of the totalstrain e:
e¼ eeþeLxs sgnðsÞ ð23Þ
where ee is the elastic strain, eL is the maximum residual strain andsgnð�Þ is the sign function. The maximum residual strain eL,regarded as a material constant, is a measure of the maximumdeformation obtainable only by multiple-variant martensite det-winning, hence, a measure of the maximum deformation obtain-able aligning all the single-variant martensites in one direction.Moreover, the presence of sgnðsÞ is the later equation indicates thatthe direction of the effect relative to the martensite fraction xs isgoverned by the stress. Finally, the elastic strain is assumed to belinearly related to the stress:
s¼ Eee ð24Þ
with E the elastic modulus.The adopted SMA model presents many advantages, such as
robustness and simplicity of algorithmic implementation asdetailed in Ref. [9].
3.3. Optimization of the cross-sectional area and the length of the SMA
wires
In the design of SMA-based damping devices to control a staycable dynamic, one of the most practical problems to be faced ischoosing the appropriate dimensions of SMA elements.
It has been observed that a cable rain/wind-induced responsetends to be dominated by the first mode or by the first few modes[34]. For simplicity, in the following analysis, we consider just onevibration mode of the cable. Then the equation of motion of thesystem is
m11 €a1 ðtÞþðk11þl2m11Þa1ðtÞ ¼ Fy1�fcðtÞj1ðxcÞ ð25Þ
In order to give an optimal parameters for the SMA, such as for thecross-sectional area and the length of wires, different criteria maybe followed. The most useful criterion design refers to energy-based methods [12]. The idea is that the SMA performs at its best ifit is capable of dissipating as much as possible of the total energy ofthe structure.
The energy balance of the equilibrium equation (25) is definedas follows:
EkðtÞþEeðtÞ ¼ EiðtÞþEcðtÞ ð26Þ
where Ek(t) is the stay cable kinetic energy defined as
EkðtÞ ¼12m11 _a1
2ðtÞ
EeðtÞ is the stay cable elastic energy defined as
EeðtÞ ¼ 12ðk11þl
2m11Þa21ðtÞ
Ei(t) is the input energy defined as
EiðtÞ ¼
Z t
0Fy1ðtÞ _a1 ðtÞ dt
luation of shape-memory-alloy superelastic behavior to control aoi:10.1016/j.ijnonlinmec.2010.12.002
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
95
97
99
101
103
105
107
109
111
113
115
119
Table 1Geometric and properties of the stay cable model.
Cable length L (m) 4.00
Mass per unit length m (kg m�1) 0.6
Inclination angle y (deg) 30
Elastic modulus E (1011N m2) 0.9
O. Ben Mekki, F. Auricchio / International Journal of Non-Linear Mechanics ] (]]]]) ]]]–]]] 5
Ec(t) is the energy associated with the SMA device defined as
EcðtÞ ¼ EecðtÞþEdcðtÞ
This energy term can be considered as the sum of an elastic term,Eec(t), and of a dissipative term, Edc(t). The dissipative term is thearea of hysteresis loop.
The optimal device in free-vibration is chosen when themaximum value of the energy dissipated in the SMA device, Edc(t),is considered. For a fixed position of the SMA, we decided tomaximize the force exerted by SMA fc(t) with the hope that thisshould load also to augmentation of dissipated energy. The forceexerted by SMA has this expression
fcðtÞ ¼ EASMAvðxc ,tÞ
LSMA�xseL signðeÞ
� �ð27Þ
where ASMA and LSMA are respectively the cross-sectional area andthe length of the SMA device, and v(xc,t) is the cable transversedisplacement at location xc. For one mode, vðxc ,tÞ ¼ a1ðtÞj1ðxcÞ.
From Eq. (27), it is clear that to maximize the value of forcedamper the cross-sectional area of the SMA device should bechosen as big as possible and the length of the SMA device should bechosen as short as passible. Therefore, the appropriate SMA devicelength should be determined by the following condition:
emaxSMA ¼ e
ASf ð28Þ
where eASf is the strain corresponds to the stress finish of the
martensite transformation sASf as indicate in Fig. 3. Thus, the
optimal length of the SMA device is
LoptSMA ¼
jvmaxðxc ,tÞj
eASf
ð29Þ
3.4. Dynamic properties and response of the combined cable-SMA
damper system
To investigate the effects of the SMA damper parameters andits location on the damping capability and control efficacy ofSMA damper, the stay cable model, which was built and which willbe studied in Civil Engineering Laboratory at the University ofTunis, was used to carry out the investigation in this study. Thegeometric and material properties of the stay cable are listed inTable 1.
The internal damping of the cable is not considered in thiswork. The stay cable is very flexible structure. However, in thisstudy, the cable is chosen such as its first modal natural frequency,f1, is equal to 1 Hz. The static tension of this stay cable can beobtained based on the identified first modal natural frequency byusing the following expression [13]:
T ¼ 4mL2f 21 ¼ 38:4 N ð30Þ
SM
A S
tress [M
Pa]
−1500
−1000
−500
0
500
1000
1500
SM
A S
tress [M
Pa]
−0.1 −0.05 0 0.05 0.1
Strain [%]
Fig. 3. Uniaxial test: numerical simulation of the complete tension–compression transfo
transformation cycles (right).
Please cite this article as: O. Ben Mekki, F. Auricchio, Performance evastay cable in cable-stayed bridges, Int. J. Non-Linear Mech. (2010), d
The free-vibration of the stay cable is examined under threedifferent scenarios to determine the effect of the SMA dampercross-sectional area, the effect of the SMA damper length, andfinally, the effect of the location of the attached point of the SMAdevice on the cable. The numerical simulation of the algorithmicmodel of the SMA is illustrated by Fig. 3 considering this data for theSMA element.
E¼ 50 000 MPa, eL ¼ 8%
sASs ¼ 500 MPa, sAS
f ¼ 600 MPa
sSAs ¼ 250 MPa, sSA
f ¼ 200 MPa
8>><>>: ð31Þ
3.4.1. Effect of the SMA damper cross-sectional area
To analyse the effect of the SMA cross-sectional area, the non-dimensional modal participation factor for the first mode of thestay cable is plotted by solving Eq. (25) and considering that theSMA damper is located at xc/L¼0.1 and the SMA length isLSMA¼230 mm chosen from Eq. (29). Fig. 4 presents the variationof the cable response according to the variation of the SMA cross-sectional area, on the left, and the variation of the SMA force, on theright. Three values of the wire diameter of the SMA damper arechosen. It is clear from this figure that the damping of the stay cabletransverse vibration increases while increasing the SMA dampercross-sectional area. This numerical study confirms the analyticalconclusion from Eq. (27) concerning the optimization of the SMAwire section.
From this simulation we can observe that the SMA damper is notable to control the small vibration because the SMA deformation issmaller than 1%.
3.4.2. Effect of the length of the SMA damper
In this example, the wire diameter of the SMA damper and thelocation of the SMA damper are fixed at DSMA¼ 0.2 mm and xc/L¼0.1,respectively, while the length of the SMA is varied as 230, 550, and750 mm. The steady-state response of the cable versus dimensionaltime is plotted on the left of Fig. 5, and, on the right, the force SMAdamper is plotted versus the displacement of the wire SMA for eachLSMA. The damping of the stay cable transverse vibration increaseswhile decreasing the length of the SMA damper. The wire SMA lengthhas more important effect on the small vibrations.
121
123
125
127
129
131
133
−0.1 −0.05 0 0.05 0.1
−1500
−1000
−500
0
500
1000
1500
Strain [%]
rmation cycle (left) and numerical simulation of the multiple tension–compression
luation of shape-memory-alloy superelastic behavior to control aoi:10.1016/j.ijnonlinmec.2010.12.002
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
95
97
99
101
103
105
107
109
111
113
115
119
121
123
125
127
129
131
133
0 5 10 15
−0.06
−0.04
−0.02
0
0.02
0.04
0.06
Time (s)
α 1 (
t)
−0.0
2
−0.0
15
−0.0
1
−0.0
05 0
0.00
50.
01
0.01
50.
02
−20
−15
−10
−5
0
5
10
15
20
f c (
t) (
N)
v (xc, t) (m)
DSMA = 0.2mmDSMA = 0.1mmDSMA = 0.05mm
DSMA = 0.2mmDSMA = 0.1mmDSMA = 0.05mm
Fig. 4. Cross-sectional area effect of the SMA damper.
0 5 10 15
−0.06
−0.04
−0.02
0
0.02
0.04
0.06
Time (s)
α 1 (
t)
−0.0
15
−0.0
1
−0.0
05 0
0.00
50.
01
0.01
50.
02
−20
−15
−10
−5
0
5
10
15
20
f c (
t) (
N)
v (xc, t) (m)
LSMA = 230mm
LSMA = 550mm
LSMA = 750mm
LSMA = 230mm
LSMA = 550mm
LSMA = 750mm
Fig. 5. Length effect of the SMA damper.
Fig. 6. Combined stay cable with TMD (left) and with SMA damper (right).
O. Ben Mekki, F. Auricchio / International Journal of Non-Linear Mechanics ] (]]]]) ]]]–]]]6
4. Comparison between SMA damper and TMD damper tocontrol the stay cable
The real uses of the control systems of the stay cable are the TMD,the cross-ties system, and the MR damper. However, the cross-tiessystem perturbs the aesthetics of the bridge and the main application ofMR is the semi-active control. For this, we choose to compare the SMAwith regard to another passive system widely used in practice.
The well known passive control energy dissipation devices instructure control is the tuned mass damper (TMD), which systemhas been widely used in civil engineering. A large number ofbuildings, such as Citicorp center, and several bridges, such asNormandie cable-stayed bridge, particularly footbridges such as
Please cite this article as: O. Ben Mekki, F. Auricchio, Performance evastay cable in cable-stayed bridges, Int. J. Non-Linear Mech. (2010), d
Millenium bridge, are now equipped with TMDs [2]. The TMDs arealso used to control stay cable in cable-stayed bridges. Tuned massdamper-magnetorheological (TMD-MR) damper for control staycable is proposed by [8,33]. Analytical, numerical, and experi-mental studies showed the efficiency of this damper and itspractical interest to control stay cable. Ref. [7] used a TMD damperto control the in-plane vibration of two stay cables. The TMD isplaced between the twin cables at their midpoints where asignificant reduction in vibration levels is notified.
In this section, the idea of the comparison between the optimalTMD and the SMA damper to control the same stay cable that wasintroduced in the previous section is investigated. The TMD deviceis considered attached at the midpoint of the cable, optimal
luation of shape-memory-alloy superelastic behavior to control aoi:10.1016/j.ijnonlinmec.2010.12.002
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
O. Ben Mekki, F. Auricchio / International Journal of Non-Linear Mechanics ] (]]]]) ]]]–]]] 7
location, as indicated in Fig. 6. In this section, for simplicity, the sagcable is not considered. The equations of the model cable-TMDsystem are written as follows:
ðm11þm�Þ €a1 ðtÞþm� €y�þk11a1ðtÞ ¼ Fy1
m� €y�þm� €a1 ðtÞþc� _y�þk�y� ¼ 0:
�ð32Þ
where m�, k�, and c� are the mass, the stiffness and the dampingcoefficient of the TMD, respectively. y� is relative displacement ofthe TMD with respect to cable.
The non-dimensional model of the system can be written as
ð1þmÞ €a1þffiffiffiffimp €Y� þa1 ¼
Fy1
m11o21
€Y �þffiffiffiffimp €a1þ2zp _Y �þp2Y� ¼ 0:
8><>: ð33Þ
with
m¼ m�
m11, o� ¼
ffiffiffiffiffiffiffik�
m�
r, o1 ¼
ffiffiffiffiffiffiffiffiffik11
m11
s
t¼o1t, _Y� ¼dY�
dt, Y� ¼
ffiffiffiffi1
m
sy�
p¼o�
o1, z¼
c�
2 m�o�
8>>>>>>>>>><>>>>>>>>>>:
ð34Þ
The non-dimensional TMD’s parameters m, p, and z are, respec-tively, related to the dimensional TMD’s ones m�, k�, and c�. Theoptimal parameters of TMD depend on the type of excitation.
95
97
99
101
103
105
107
109
111
0 2 4 6 8 10
−0.06
−0.04
−0.02
0
0.02
0.04
0.06
Time (s)
α 1 (
t)
With optimal TMD
With SMA
Fig. 7. Comparison between SMA damper and TMD to control the stay cable free
vibration.
0 2 4 6 8 10
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
Time (s)
α 1 (
t)
With optimal TMDWith SMA
Fig. 8. Comparison between SMA damper and TMD to control the stay cable harmonic vib
with SMA damper (right).
Please cite this article as: O. Ben Mekki, F. Auricchio, Performance evastay cable in cable-stayed bridges, Int. J. Non-Linear Mech. (2010), d
However, the optimal design of the TMD for a fixed ratio of the TMDmass for free vibration [27] are
poptm ¼
1
1þm
zoptm ¼
ffiffiffiffiffiffiffiffiffiffiffim1þm
r8>>><>>>:
ð35Þ
while for an harmonic excitation [19] they are
poptm ¼
1
1þm
zoptm ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3m
8ð1þmÞ
s8>>>><>>>>:
ð36Þ
In Fig. 7 the free-response of the stay cable is plotted when it iscontrolled by TMD and SMA damper, solving, respectively, system(32) and system (25) without external force. The TMD is attached inthe midpoint of the cable and these optimal parameters areobtained from system (35) considering m¼ 5%. On the other hand,SMA damper is placed near the end with xc/L¼0.1. This position isnot optimal but it is the most practical position in reality. Thelength and the cross-sectional area of SMA damper are fixed fromthe parametric study made in the previous section. These para-meters took the optimal values: LSMA¼230 mm and DSMA¼0.2 mm.
Fig. 7 shows that the SMA, in its non-optimal position, is able todamp only the high free vibration of the stay cable better than theoptimal TMD in its optimal position. However, it is able to damp theharmonic excitation more better than the optimal TMD as shown inFig. 8. The parameters of the SMA are always the same whereas theoptimal parameters of the TMD are obtained in this case accordingto formula (36).
From Fig. 7, we can observe that the largest vibration response ofthe SMA dampers controlled cable is around 50% of the oneobtained using the TMDs, and the vibration decay speed of theSMA controlled cable is much faster than that of the TMD controlledcable, suggesting that the proposed SMA damper is very effective toreduce structural response. An equivalent viscous damping value ofthe TMD is 7.3% however the equivalent viscous damping value ofthe shape-memory alloys is 17.45% when we consider only thevibration response before the stabilization of the oscillation (2 s).
The only one favors of TMD, in free vibration, is able to betterdamp out the small vibrations.
From the right of Fig. 8 it is observed that without any dampingdevice the cable vibration amplitude is much larger than when theSMA damper is installed.
113
115
119
121
123
125
127
129
131
133
0 2 4 6 8 10
−4
−3
−2
−1
0
1
2
3
4
Time (s)
α 1 (
t)
With optimal TMDWith SMA
ration (left) and harmonic vibration of the stay cable first mode without actuator and
luation of shape-memory-alloy superelastic behavior to control aoi:10.1016/j.ijnonlinmec.2010.12.002
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
95
97
99
101
103
105
O. Ben Mekki, F. Auricchio / International Journal of Non-Linear Mechanics ] (]]]]) ]]]–]]]8
5. Conclusions
A numerical study on the oscillation mitigation of a combinedcable-SMA damper system has been conducted. The purpose of thisinvestigation is to devise a SMA damper with superelastic proper-ties to eliminate most types of oscillations and to understand theoptimal parameters such as section and length of the SMA.
The simulation on a cable model with SMA damper showed theexcellent property of energy dissipation of the SMA and their abilityto suppress the cable vibration. The performance of the SMAdamper is notably dependent on the cross-sectional area, thelength and the position of the SMA. The cross-sectional area andthe length of the SMA device should be chosen, respectively, as bigas possible and as short as possible. Finally, the comparison fromthe SMA damper and the well known passive control energydissipation devices in structure control, the TMD, showed thatthe SMA, in its non-optimal position, is able to damp the high freevibration of the stay cable better than the optimal TMD, in itsoptimal position, and to damp the harmonic excitation more betterthan the optimal TMD.
References
[1] B. Andrawes, R. DesRoches, Sensitivity of seismic application to different shapememory alloy models, Journal of Engineering Mechanics 183 (2008) 134–173.
[2] B.F. Spenser, S. Nagarajaiah, State of the art of structural control, Journal ofStructural Engineering (ASCE) 129 (7) (2003) 845–856.
[3] B.M. Pacheco, Y. Fujino, A. Sulekh, Estimation curve for modal damping in staycable with viscous damper, Journal of Structural Engineering (ASCE) 119 (9)(1993) 1961–1965.
[4] C. Cremona, Dynamique non lineaire des cables et application au comporte-ment des ponts haubanes, Revue Franc-aise de Mecanique 2 (2000) 107–113.
[5] C. Cremona, A generalisation of the Pacheco’s curve for linear damper design.in: Proceedings of the 4th International Symposium on Cable Dynamics, Liege,A.I.M, 2001, pp. 293–308.
[6] C. Cremona, J.C. Foucriat, Comportement au vent des ponts, Presses EcoleNationale des Ponts et Chaussees (ENPC), 2002.
[7] C.N. Jensen, Optimal damping of stays in cable-stayed bridges for in-planevibrations, Journal Sound and Vibration 256 (2002) 499–513.
[8] C.S. Cai, M. Asce, W.J. Wu, M. Araujo, Cable vibration control a TMD-MR dampersystem: experimental exploration, Journal of Structural Engineering (ASCE)133 (2007) 629–637.
[9] F. Auricchio, E. Sacco, A one-dimensional model for superelastic shape-memory alloys with different elastic properties between austenite andmartensite, International Journal of Non-Linear Mechanics 32 (6) (1997)1101–1114.
[10] F. Grandhi, D. Wolons, Characterization of the pseudoelastic damping behaviorof shape memory alloy wires using complex modulus, Smart Materials andStructures 8 (1999) 49–56.
[11] G. Song, N. Ma, H.N. Li, Applications of shape memory alloys in civil structures,Engineering Structures 28 (2006) 1266–1274.
[12] G.W. Housner, L.A. Bergaman, T.K. Caughey, A.G. Chassiakos, R.O. Claus,S.F. Masri, R.E. Skelton, T.T. Soong, B.F. Spencer, J.T.P. Yao, Structural control:past, present, and future, Journal of Engineering Mechanics 123 (9) (1997)897–971.
[13] H.M. Irvine, Cable Structures, The MIT Press, Cambridge, 1981.[14] H.N. Li, H. Qian, G.B. Song, D.W. Gao, Type of shape memory alloy damper:
design, experiment and numerical simulation, Journal of Vibration Engineering21 (2008) 179–184.
[15] I. Kovacs, Zur frage der seilschwingungen und der seildampfung, Die Bau-technik 59 (1982) 325–332.
Please cite this article as: O. Ben Mekki, F. Auricchio, Performance evastay cable in cable-stayed bridges, Int. J. Non-Linear Mech. (2010), d
[16] J. Ocel, R. DesRoches, R.T. Leon, J.R. Hayes, S. Sweeney, Steel beam-columnconnections using shape memory alloys, (ASCE) Journal of Structural Engi-neering 130 (2004) 731–740.
[17] J.A. Main, S. Jones, Evaluation of viscous dampers for stay-cable vibrationmitigation, Journal of Bridge Engineering (ASCE) 6 (6) (2001) 385–388.
[18] J.A. Main, S. Jones, Free vibrations of taut cable with attached damper.Part i: linear viscous damper, Journal of Engineering Mechanics 128 (2002)1062–1071.
[19] J.P. Den Hartog, Mechanical Vibrations, McGraw Hill Book Co., New York, 1947.[20] K.H. Ip, Energy dissipation in shape memory alloy wire under cyclic bending,
Smart Materials and Structures 9 (2000) 653–659.[21] L. Hui, L. Min, O. Jinping, Vibration mitigation of stay cable with one shape
memory alloy damper, Structural Control and Health Monitoring 11 (2004)21–36.
[22] M. Arata, N. Mohammad, Optimization of hysteretic characteristics of dampingdevices based on pseudoelastic shape memory alloys, International Journal ofNon-Linear Mechanics 37 (2002) 1375–1386.
[23] M. Dolce, D. Cardone, Mechanical behaviour of shape memory alloys forseismic application 1. Martensite and austenite NiTi bars subjected to torsion,International Journal of Mechanical Sciences 43 (2001) 2631–2651.
[24] M. Indirli, M.G. Castellano, P. Clemente, A. Martelli, Demo application of shapememory alloy devices: the rehabilitation of S. Georgio Church Bell Tower, in:Smart Systems for Bridges, Structures and Highways, SPIE, vol. 4330, 2001,pp. 262–272.
[25] M.D. Symans, M.C. Constantinou, Semi-active control systems for seismicprotection of structures: a state-of-the-art review, Engineering Structures 21(6) (1999) 469–487.
[26] N. Gimsing, Cable-stayed bridges-past, present, and future. in: ProceedingsIABSE, 1999.
[27] O. Ben Mekki, F. Bourquin, F. Maceri, C. Nguyen, An adaptive pendulum forevolving structures, in: Proceedings of the 4th IASC World Conference onStructural Control and Monitoring, San Diego, CA, July 2006.
[28] R. DesRoches, M. Delemont, Seismic retrofit of simply supported bridges usingshape memory alloys, Engineering Structures 24 (2002) 325–332.
[29] R.W. Poston, Cable-stay conundrum, ASCE Civil Engineering 768 (8) (1998)58–61.
[30] T. Susumpow, Y. Fujino, Active control of multi-modal cable vibrations byaxial support motion, Journal of Engineering Mechanics (ASCE) 121 (1995)964–972.
[31] V. Torra, A. Isalgue, C. Auguet, F. Martorell, P. Terriault, F.C. Lovey, Built indamping for family homes via SMA: an ANSYS computation scheme based onmesoscopic and microscopic experimental analyses, Engineering Structures 29(2007) 1889–1902.
[32] V. Torra, A. Isalgue, C. Auguet, G. Carreras, F.C. Lovey, H. Soul, P. Terriault,Damping in civil engineering using SMA. The fatigue behavior and stability ofCuAlBe and NiTi alloys, Journal of Materials Engineering and Performance 18(2009) 738–745.
[33] W.J. Wu, C.S. Cai, Theoretical exploration of a taut cable and a TMD system,Engineering Structures 29 (2007) 962–972.
[34] X.Y. Wang, Y.Q. Ni, J.M. Ko, Z.Q. Chen, Optimal design of viscous dampers formulti-mode vibration control of bridge cables, Engineering Structures 27(2005) 792–800.
[35] Y. Fujino, P. Warnitchai, B.M. Pacheco, Active stiffness control of cablevibration, Journal of Applied Mechanics (ASME) 60 (1993) 948–953.
[36] Y. Liu, Z. Xie, J.V. Humbeeck, Cyclic deformation of NiTi shape memory alloys,Materials Science and Engineering A 273–275 (1999) 673–678.
[37] Y.L. Han, Q.S. Li, A.Q. Li, A.Y.T. Leung, P.H. Lin, Structural vibration control byshape memory alloy damper, Earthquake Engineering and StructuralDynamics 32 (2003) 483–494.
[38] Y.L. Xu, H.J. Zhou, Damping cable vibration for a cable-stayed bridge usingadjustable fluid dampers, Journal of Sound and Vibration 306 (2007)349–360.
[39] Y.Q. Ni, X.Y. Wang, J.M. Ko, Z.Q. Chen, Field observations of rain–wind-inducedcable vibration in cable-stayed Dongting Lake Bridge, Journal of WindEngineering and Industrial Aerodynamics 95 (2007) 303–328.
[40] Z. Yu, Y.L. Xu, J.M. Ko, A hybrid method for estimating modal damping in saggedcable with damper, Journal of Advances in Structural Engineering 1 (1998)212–236.
107
luation of shape-memory-alloy superelastic behavior to control aoi:10.1016/j.ijnonlinmec.2010.12.002